Pricing
Plug your agents in. Plug your team in. Plug Claude in. Plug your custom agent in. Anything that reads or writes the customer memory graph is welcome. We don't count readers — we charge for accounts tracked and capabilities unlocked.
Unlimited readers — agents + humans
For 1–50 person CS teams launching their first AI agents.
Unlimited readers — agents + humans
For teams running real account portfolios with custom agents.
Unlimited readers — agents + humans
For multi-product, multi-org companies with stricter requirements.
The pricing principle
Hyperengage was designed for an agent-native post-sales world — but agents and humans use the same graph, with the same scoped tokens, against the same API. We don't care whether the reader is your VP of CS or a Claude agent drafting a renewal email. Plug it in, and pay for the graph.
Plug in your agents
Sidekick, Copilot, Signals Engine, MCP-connected runtimes (Claude, Codex, ChatGPT), your own custom agents, batch jobs, scripts. Anything that speaks MCP can read and write — at any rate, in parallel, around the clock.
Plug in your team
CSMs, AEs, RevOps, PMs, execs. Whoever steers strategy gets to read the graph through the UI, override what agents propose, and write decisions back. No seat tax, no "AI module" upcharge.
We charge for the graph
Pricing follows the asset (one Customer Memory Graph), the unit of value (accounts tracked), and the capabilities you unlock (MCP, custom skills, advanced integrations). Not the count of readers.
A Monday on Hyperengage
·memory.activity_log
decision nodes8 draftsSame graph. Same tokens. Nobody counts logins.
Feature matrix
Starter $500/mo | Scale $2,000/mo | Enterprise Custom | |
|---|---|---|---|
Customer Memory Graph | |||
| Customer Memory Graphs | 1 | 1 | Unlimited |
| Customer accounts tracked | 250 | 2,500 | Unlimited |
| Memory entity types | All 7 | All 7 | All 7 + custom |
| Reconciliation across sources | |||
| Bidirectional reads + writes | |||
| Data retention | 90 days | 1 year | Custom |
| Audit log | 30 days | 1 year | Custom |
Readers (agents + humans) | |||
| Total readers — agents + humans combined | Unlimited | Unlimited | Unlimited |
| Sidekick — CSM briefings | |||
| Copilot — natural-language Q&A | |||
| Signals Engine — playbook automation | |||
| MCP API (Claude, Codex, any agent) | |||
| Custom agent skills | 5 | Unlimited | Unlimited |
| Bring-your-own-LLM (Anthropic, OpenAI, Vertex, Bedrock) | |||
| Reader rate limits | Standard | 10× standard | Custom |
Integrations | |||
| HubSpot, Salesforce | |||
| Segment + product events | |||
| Intercom, Zendesk | |||
| Slack, email | |||
| Stripe, Chargebee | |||
| Custom webhooks | |||
| Salesforce CPQ + custom CRM | |||
| Reverse ETL (Snowflake, BigQuery) | |||
Access & identity | |||
| Roles & permissions | Standard | Custom | Custom + audit |
| SSO / SAML | |||
| SCIM provisioning | |||
| Sandbox environment | |||
| Per-reader scoped tokens | |||
Security & compliance | |||
| SOC 2 Type II | In progress | In progress | In progress |
| GDPR + DPA | |||
| PII redaction | |||
| Per-workspace logical isolation | |||
| Single-tenant infrastructure | |||
| Penetration test reports | |||
| Custom security review | |||
Support & SLA | |||
| Email support | |||
| Shared Slack channel | |||
| Priority response | Same business day | 24/7 | |
| Uptime SLA | 99.5% | 99.9% | 99.99% |
| Quarterly business review | |||
| Dedicated CSM | |||
| White-glove onboarding | |||
Agent-native architecture
Every Hyperengage capability is exposed as a typed MCP tool. Your agents read the graph the same way your team does — just a thousand times faster, around the clock, in parallel.
// any reader — agent or human-attributed — uses the same API
const brief = await memory.read({
account: "acme.com",
since: "last_touch",
include: ["events", "decisions", "signals"],
});
// write a decision back — attributedTo can be agent or user
await memory.write({
type: "decision",
summary: "Renewal: hold pricing, offer 12mo term",
attributedTo: "agent.claude",
});
Every reader is typed
Each reader (agent or human) gets a scoped token, identity, and audit trail. attributedTo lets you trace every read and write back to the source.
MCP-native, model-agnostic
Claude, Codex, ChatGPT, Cursor, Continue, your own runtime, your own scripts. Anything that speaks MCP plugs into the same graph.
Bidirectional, not read-only
Readers don't just consume — they write decisions, actions, and signals back into the graph for the next reader to see.
Built into every plan
SOC 2 in progress
In active audit window. Posture & timeline on request.
GDPR + DPA
Standard DPA on every plan.
Unlimited readers
Agents + humans. Bring as many as you want.
Global deployment
Hosted in the US with Enterprise options.
Your data is yours
Full export. No lock-in. No retraining.
All standard integrations
HubSpot, Salesforce, Segment, Slack, Stripe.
Real human support
Slack channel from day one.
Free 14-day trial
No credit card. No sales call required.
We were YC's ICP before we were Hyperengage. If you're building a B2B SaaS company and your CS team needs an agent-native memory graph from day one, we want you on the platform — at a price that respects pre-PMF realities.
$250/month for 12 months
Full Starter — unlimited readers, 3 built-in agents, 250 accounts. No pro-rate gotchas.
Founder office hours
Every other Friday with our founders. Bring your retention dashboard.
MCP at Starter pricing
We unlock the MCP API on Starter for the program duration.
No long-term contract
Month-to-month. If we don't earn it every month, we don't deserve it.
Eligibility: Active or alumni YC / Techstars batch, OR pre-Series A with under $5M ARR.
FAQ
Because everything that reads the memory graph — agents, humans, scripts, integrations — is just a reader. Sidekick, Copilot, Signals Engine, Claude (via MCP), your custom agents, your CSMs, your AEs, your execs: plug them all in. We don't count. We charge for the graph (the asset) and the accounts you track (the unit of value). Bring every reader.
No. We don't differentiate. Anything that plugs into the memory graph is a reader, full stop. Pricing is the same whether you have 5 humans and 50 agent calls/day, or 200 humans and 5,000 agent calls/day. The work happens against the same graph.
A customer account is one company you're tracking in your memory graph. If you have 200 enterprise customers and 50 SMB customers, that's 250 accounts. Internal users, prospects, and former customers don't count.
We flag it within 7 days and reach out — no auto-charges, no rate limits on your existing graphs. You can either scope down or upgrade. Most teams hit the next tier when they're ready, not because we forced it.
MCP (Model Context Protocol) is the open standard for letting AI agents call typed tools. Hyperengage exposes the Customer Memory Graph as an MCP server, so any MCP-compatible reader — Claude, Codex, ChatGPT desktop, your own runtime — can read live customer state and write decisions back. Available on Scale and Enterprise.
Yes, on Enterprise. Connect Claude (Anthropic), OpenAI, Vertex AI (Google), or Bedrock (AWS). Your data stays in your provider's account; we never co-mingle inference with another tenant.
Yes. Every plan starts with a 14-day free trial. No credit card required. Connect your stack, build your first memory graph, and let your readers — agent and human — use it before you decide.
Yes — 50% off Starter for the first 12 months, plus founder office hours every other week. See the Hyperengage Startup Program above. We accept any active or alumni YC company at any batch.
Most teams have a working memory graph in 48 hours. Connect your CRM, product analytics, and Slack; we reconcile the rest. No 6-week professional services engagement, no $25K setup fee.
Yes. Monthly plans cancel at the end of the cycle, no questions asked. Annual plans pro-rate the unused portion if you need to exit early. No auto-renewal traps.
SOC 2 Type II is in progress — we're partway through the audit window with our auditor. We can share our current security posture, controls documentation, and timeline on request. GDPR + DPA, PII redaction, and per-workspace logical isolation are already in place. Single-tenant infrastructure is available on Enterprise.
Yes, on Enterprise — SSO/SAML and SCIM provisioning are included. Custom security review, custom MSA/DPA, and custom contract terms come standard with the Enterprise plan.
Connect your stack. Plug every reader in — agents and humans. Pick a plan when you're ready. Cancel anytime — though most teams stay because the graph keeps getting smarter.
No credit card required · Unlimited readers — agents + humans